## Thursday, December 09, 2010

### The utter balls people write about Oxbridge

I start by apologizing for the profanity, but when I hear people spouting questions like "What is it about the famed Oxbridge interview system that counts against students who didn't attend a top public school?" it makes me very angry. The implication in the article is that there's a race bias (or is it a school bias, or a north/south divide bias... I actually lost track of the number of biases the article is claiming).

The answer to the question is rather simple (as long as the question is reframed as "What is it about the famed Oxbridge interview system that counts against some people?"). The Oxford interview process is bloody hard. So hard that 25 years on I use questions asked of me at 17 years old to screen candidates for programming jobs. The interview process is not designed to discriminate against people who didn't go to a top public school, it's meant to discriminate against people who aren't up to studying there.

I attended Oxford at the same time as David Cameron and his chums (Michael Gove was in the room next to me for a year). There certainly were lots of people from public schools (perhaps they did get in because their Dad went to Oxford, or perhaps it was because of the level of education they received), but there were also lots of people like me who didn't, in the stereotype, go to Eton.

I went to Oxford from a large comprehensive school. I sat that grueling entrance exam in mathematics, I was invited for interview and stayed days in Oxford being interviewed over and again. I didn't get special tuition to make it into Oxford, I'm not a public school boy and no one in my family has an Oxford connection. Neither of my parents have degrees.

I was asked extremely searching questions about mathematics and computer science that were well outside any A level curriculum and the purpose was to see how I would think. One interviewer pointedly asked me why I hadn't done a particular question on the entrance exam and then made me answer it at a blackboard in front of him. Another made me stand in front of a blackboard and solve a problem in computer science.

While I was at Oxford I was asked to go into comprehensive schools to encourage people to apply. Many people write themselves off and don't even try. This is a problem and the linked article doesn't help the situation by portraying Oxford as racist.

The author should have asked himself why so few black students were applying to Oxford and so few were getting top A level grades. You'd think he might have done that given that he was Minister for Higher Education under the previous government. But it's a lot easier to point the finger at some imagined evil institution than ask the hard questions about the state of education in British schools.

And he really shows his deep knowledge of the subject when he states: "Cambridge doesn't employ a single black academic." Sorry, Dr Okeoghene Odudu, Dr Justice Tankebe (inter alia) I guess you don't count for some reason.

It is tragic that such a small number of black students are getting top grades, but whacking Oxford and Cambridge without attacking the root cause is almost criminal. It's betraying the people the author wishes to be believed to be trying to help.

He also states: "You will not find these figures on the Oxford or Cambridge websites. ". Wanna bet? How about Oxford's Undergraduate Admissions Statistics 2009 entry and let's look at Ethnic Origin.

And we'll just compare "Whites" to "Black African, Caribbean and Other". So 8,378 white applicants; 221 black. Swap to acceptances we have 2,316 white acceptances; 27 black. So 28% of white applicants got in and 12% of black. Evidence of race bias or something else?

To quote the site: "Oxford’s three most oversubscribed large (over 70 places) courses (Economics & Management, Medicine and Mathematics) account for 44% of all Black applicants – compared to just 17% of all white applicants." and "Subject breakdown: 28.8% of all Black applicants for 2009 entry applied for Medicine, compared to just 7% of all white applicants. 10.4% of all Black applicants for 2009 entry applied for Economics & Management, compared to just 3.6% of all white applicants."

So you've got a small number of candidates applying into the most oversubscribed subject areas. 44% of black applicants are applying for courses with acceptance rates of 7.9%, 12.1% and 19%.

Put those figures together and assume no bias and for the 44% you've got a 12 black students who get admitted out of 97 who apply to those subjects. That's a success rate of 12%. That gibes with the figure given above: and that's assuming that the acceptance rate for those three subjects has no bias at all.

What about the other 56%? That's 123 students of which 27 - 12 got accepted. So that's also 12%. The problem with interpreting that is those 15 students are a tiny portion of the pool of 11,896 students who applied. And without knowing what subjects they applied for it's hard to dig into them.

But it is possible to work backwards. Suppose that 28% of those 123 black students were accepted (the average rate for whites) then there'd be 34 accepted. So the total would be 34 + 12 out of 221 or 20.8%. Comparing that with the overall 12% rate it's clear that the acceptance rate for black students is lower than white students in the non-oversubscribed subjects. But knowing why is hard.

If they are all applying for earth sciences (acceptance rate 44.9%) then there's a problem, if they are applying for law (acceptance rate 17.7%) then a different picture emerges. And if it's Fine Art (acceptance rate 12.9%) they are close to spot on. The only way to the bottom of that puzzle is a breakdown by subject and ethnic origin. But with such a tiny group of applicants even a change of acceptance of a single student could cause wild swings in percentage acceptance rates.

The other laughable misuse of statistics in the article come in the form of cherry-picking. "Merton College, Oxford, has not admitted a single black student for five years." Hardly surprising. If 2009 isn't anything to go by just 27 black students were admitted to the entire university. There are 38 colleges in Oxford. It's not possible to divide 27 by 38 evenly and no surprise that a specific college would have no black student for a number of years.

The bottom line is that getting into Oxbridge is hard and that the number of black students applying is tiny. Imagine for a moment that black students got in at the same rate as white students. There would still only be 62 black students at the university. Let's attack the real problem and raise up the education level of black students.

Update: Follow up post looking into the Merton Problem.

Update: I emailed Oxford asking if they'd release the breakdown by ethnic origin and subject so that per-subject bias can be examined in the non-over subscribed subjects. Will blog if I get a result.

Update: I saw a comment on Twitter that said that it was "delusional of the author [i.e. me] to doggedly say there is no way oxbridge has any institutional issues at all." Clearly, I haven't said that Oxford has no institutional issues (in fact, it would be utterly amazing if it didn't), and in a comment I stated: "If anyone would like to point me to statistically significant data that shows bias I'd be happy to write about it." I don't see it from the data, but if it's there it should be examined.

Labels:

If you enjoyed this blog post, you might enjoy my travel book for people interested in science and technology: The Geek Atlas. Signed copies of The Geek Atlas are available.

<$BlogCommentBody$>

<$BlogCommentDateTime$> <$BlogCommentDeleteIcon$>

<$BlogBacklinkControl$> <$BlogBacklinkTitle$> <$BlogBacklinkDeleteIcon$>
<$BlogBacklinkSnippet$>